Website Content Crawler avatar
Website Content Crawler

Pricing

Pay per usage

Go to Store
Website Content Crawler

Website Content Crawler

apify/website-content-crawler

Developed by

Apify

Maintained by Apify

Crawl websites and extract text content to feed AI models, LLM applications, vector databases, or RAG pipelines. The Actor supports rich formatting using Markdown, cleans the HTML, downloads files, and integrates well with 🦜🔗 LangChain, LlamaIndex, and the wider LLM ecosystem.

4.6 (38)

Pricing

Pay per usage

1.1k

Monthly users

5.9k

Runs succeeded

>99%

Response time

2.3 days

Last modified

6 days ago

DM

Poor results

Closed
Digital_Mole opened this issue
3 months ago

I can't seem to get good results from simple documentation sites like OpenAI - most pages return partial info

jiri.spilka avatar

Hi, I apologize for the delayed response.

Unfortunately, I've tried to resolve the issue, but was not very successful. I’ve reached out internally, and we’ll try to figure this out.

Thank you for your understanding. Jiri

jiri.spilka avatar

Hi, thank you for your patience, and I apologize for the earlier delay in responding.

Credit to @jindrich.bar for this analysis

It appears the OpenAI documentation website behaves a bit unusually. However, he has identified some adjustments that can help improve the results:

  1. Ensure content is fully loaded: Use "waitForSelector": ".anchor-heading" to make sure there’s content on the page before extraction begins.
  2. Wait for dynamic content: Set "dynamicContentWaitSecs": 30" to allow sufficient time for the content to load.
  3. Avoid dropping any content: Use "htmlTransformer": "none" to prevent content from being stripped during processing.

Here’s an example run for the OpenAI documentation where these settings were applied: example run.

Additionally, he noticed that many URLs in the OpenAI docs point to different sections of the same page (e.g., the API reference). As a result, the extracted results may include the entire API reference text multiple times. To optimize costs and performance, you may want to reduce the number of URLs being processed, as the current approach with a 30-second wait can become expensive.

Please let me know if you’d like further clarification. Jiri

jiri.spilka avatar

I’ll go ahead and close this issue for now. However, feel free to ask additional questions or create a new issue if needed.

Pricing

Pricing model

Pay per usage

This Actor is paid per platform usage. The Actor is free to use, and you only pay for the Apify platform usage.